Drivers of herbivore diversity decoupled by leveraging the fossil record

Understanding the origin and maintenance of diversity is a fundamental challenge in macroecology (1). The relationship between narrow niche breadth and species richness in a community is viewed as both an explanation for coexistence through specialization and a consequence of it (1, 2). However, this relationship is further complicated by the overall breadth of niche availability across all species in the community (2). A notable example is the high diversity of phytophagous insects, particularly in relation to their dietary niche (diversity of host plants) within a community (3–5). In PNAS, Albrecht et al. (6) address this crucial issue from a functional ecological standpoint by utilizing a fossil herbivory dataset and disentangling the impacts of niche partitioning (specialization) from niche availability (host plant diversity).

Plants and their insect herbivores dominate terrestrial ecosystems in terms of both biomass and diversity today, and this has been the case for the past 420 My (today, more than one-fourth of all known multicellular species are herbivorous insects) (4, 7). The elevated intensities of their coevolutionary interactions have resulted in their unprecedented diversity (8). However, the exact reasons behind the high diversity of phytophagous insects remain fuzzy (2, 4). There are two main hypotheses: One suggests that it arises from abundant niche availability due to diverse host plants, while the other suggests that it results from intense niche partitioning among herbivores on different hosts through eco-evolutionary processes (2, 4, 9). However, most previous studies focused on either host diversity or partitioning independently and within the modern ecological context, despite the importance of both factors and the long-term coevolutionary dynamics between plants and their herbivores in shaping this diversity (3, 4). This long-term evolutionary history is afforded in great detail through the fossil record and is essential for understanding the structure of extant plant–insect interactions (7, 10). Paleontological perspectives are particularly valuable as modern organismic traits are often not representative of the past, and fossil organisms frequently have no modern counterparts; therefore, the record also represents non-analog and no-analog communities (11–13).

Albrecht et al., provide new insights into the processes behind high (phytophagous) insect diversity using a quantitative framework based on plant-DT association networks.

Herbivory in the fossil record is usually in the form of damage inflicted by herbivores on plant parts, especially foliage. These feeding damages can be classified into distinct damage types [DTs; (14)], corresponding to insect herbivore diversity in fossil and modern ecosystems (15). These distinct patterns of feeding (DTs) can further be classified into functional feeding groups (FFGs) which represent larger guilds groups of insects (e.g., hole feeders, margin feeders, skeletonizers, piercers and suckers, miners, gallers, etc.) (14). These DTs can therefore be seen as functional traits of herbivory, which have remained consistent across geographical space and evolutionary time, due to repeated convergences in mouthparts and feeding strategies among phytophagous insects (16, 17).

Previous work has extensively looked at diversity-generating and maintaining processes in both extant and extinct insect lineages as well as their relationship with host availability and diversification (4). While these studies have yielded invaluable insights into their diversity, they have all focused on a taxonomic, or species-based, perspective (4). Using a functional (or trait-based) perspective, in place of a purely taxonomic one, has shown the importance of ecological drivers on interactions in both the modern and paleo-ecological literature and provides a complementary perspective to species-based studies (16, 18–20). In the case of DTs and their host plants, the rich fossil record has allowed the exploration of important and longstanding ecological questions across large spatiotemporal scales (10, 21–24). It has also allowed for providing novel perspectives on ecological and evolutionary questions, such as the impact of environmental factors, spatiotemporal heterogeneity, and major geological events (such as mass extinctions) in shaping these plant–insect interactions (7, 10, 21–24). Continuing that tradition, in PNAS, Albrecht et al. (6) look at the relative importance of host diversity and niche packing in shaping herbivore diversity in this issue.

By going beyond the species-based paradigm and looking collectively at the relative importance of niche partitioning (related to diversity-generating and maintaining processes) and total niche breadth (host availability) across 21 European Cenozoic sites (covering almost the last 66 My), Albrecht et al. (6) provide new insights into the processes behind high (phytophagous) insect diversity using a quantitative framework based on plant–DT association networks (Fig. 1) (10, 11, 22, 25). By fitting a general model of diversity to these association networks from fossil assemblages, they find that host partitioning among the DTs (functional groups of herbivores) contributed twice as much to their functional diversity as host plant diversity (6). To address spatiotemporal autocorrelation in the fossil record, the study explicitly accounted for each fossil flora’s geographic location and age in their modeling approach—which lent more credence to their results.

Fig. 1.

Factors impacting the diversity of herbivory (DTs) through the lens of a host plant–DT network.

The relative importance of host partitioning and availability is usually explored in the modern through the means of the latitudinal biodiversity gradient, often specifically focusing on the high diversity in tropical rainforests (3–5). As a proof of concept, Albrecht et al. (6) fit their model to modern insect herbivore communities in tropical rainforests and validated their original (fossil record-based) findings, suggesting general applicability to modern ecosystems. The results from this work also provide an explanation for some previously reported intriguing phenomena such as the decoupling of the plant and insect functional diversity after the K-Pg mass extinction (22) and reduced specialized herbivory during parts of the Cenozoic (23, 24). For example, a 64.4-My-old low-diversity flora in southeastern Montana displays high herbivory diversity, particularly for leaf-mining, whereas an extremely diverse 63.8-My-old flora in the Denver Basin shows minimal damage and lacks specialized feeding (22). Similarly, an observed decline in functional diversity of specialized herbivores during the Eocene–Oligocene Transition and Late Miocene may have been associated with reduced host partitioning, likely due to disruptions in plant–herbivore associations caused by global cooling trends and significant vegetation changes during those times (6, 23, 24).

The authors suggest three general functional processes that impacted niche packing during the Cenozoic (which are also applicable to modern assemblages): 1) host plant traits, such as nutrient stoichiometry and palatability, 2) coevolutionary processes involving physiological and morphological traits of both hosts and herbivores, which may impact host dispersal, herbivore specialization, and diversification in plants and insects, 3) neutral, stochastic ecological processes diverging local functional communities on a single plant host across large spatiotemporal scales, especially given the scale difference in life history processes of plants and insects (Fig. 1) (4). This study paves the way for further exploring questions posed from a trait-based perspective, testing these explanations in both modern and paleo-ecological contexts, and delving into the intricacies of local context dependence within a broader ecological framework. The study also highlights the potential for a meaningful dialogue between modern and paleo-ecologists, enabling the examination of macroecological theories using fossil data. Despite the sparser availability of fossil records, they offer a unique perspective that extends beyond the longer temporal resolution they provide.